基于EGA算法的小样本非线性残差灰色Verhulst计量组合预测模型

针对复杂经济系统下时间序列所呈现出的小样本非线性残差特征,采用非线性残差灰色Verhuilst模型进行研究,修正传统计量模型对于残差信息挖掘不够,预测精度不高的问题,在此基础上,选择带有精英策略的EGA算法来建立灰色Verhulst计量组合预测模型,设计了算法实现的逻辑流程和非线性残差灰色Verhulst计量组合预测模型的整体建模思路,提出了改进多准则目标优化NP完全问题的新方法,对模型的预测效果进行比较分析.实证研究表明:基于EGA算法的小样本非线性残差灰色Verhulst计量组合预测模型算法收敛速度快,拟合效果好,预测结果更精确.

Study on nonlinear residual grey Verhulst metering combination forecasting model of small sample based on EGA algorithm

In the condition of complex economic system for the nonlinear residual characteristics of small sample time series, the paper chooses nonlinear grey Verhulst model to solve the problem for less residual information and not high prediction accuracy by correcting traditional econometric model. On this basis, we choose EGA algorithm with elitist strategy to establish grey Verhulst metering combination forecasting model, design the modeling algorithm logic flow and overall idea of nonlinear residuals grey Verhulst combination forecasting model, provide a new method to improve the multi-objective optimization criteria NP-complete problems, compare the forecasting model results with other models. Empirical studies show that:the nonlinear algorithm EGA residuals grey Verhulst metering combination forecasting model of small sample based on EGA algorithm converges fast, with good fitting results and more accurate predictions.